Recent in vitro work and a short clinical study suggest that adding a bonding agent layer between sealant and saliva-contaminated enamel allows for adequate bond strength and retention of resin sealants and may improve success of all sealant applications. This five-year clinical study scored 617 occlusal and 441 buccal/lingual molar sealants, with use of a split-mouth design, with half receiving sealant alone and half bonding agent plus sealant. Treatment effects and potential risk factors for sealant failure were tested by means of a Cox regression model. Three bonding agent groups were analyzed for treatment effect: Tenure primer, Scotchbond Multi-Purpose, and 3 single-bottle dentin bonding agents as a third group. The single-bottle group was successful in reducing risk of sealant failure, with a hazard ratio (HR) of 0.53 (p = 0.014) for occlusal and 0.35 (p = 0.006) for buccal/lingual sealants. Scotchbond was detrimental to occlusal sealant success, with a HR of 2.96 (p = 0.0003). Tenure primer was neutral, showing HRs close to 1.0. Variables that affected success differed between occlusal and buccal/lingual sealants, suggesting that failures on these two surfaces may be dependent upon differing factors. Early eruption stage was a significant risk factor for both surfaces (HR = 2.91, p = 0.00001, occlusal; and HR = 1.52, p = 0.015, buccal/lingual). Behavior (HR = 1.96, p = 0.0007), salivary problems (HR = 1.73, p = 0.002), and visually apparent variations in enamel (HR = 1.51, p = 0.018) were significant risk factors for occlusal sealants only. In addition to completing detailed analyses of risk factors for sealant survival, this study shows that single-bottle bonding agents protect sealant survival, yielding half the usual risk of failure for occlusal sealants and one-third the risk of failure for buccal/lingual sealants.